Insights
Questions that come before the project
Straight answers to the questions people ask when deciding where to use data and AI, and who to work with. Each piece starts from a real, anonymized problem.
AI that decides
How much of your customer service can you hand to an AI agent?
Automating customer service with AI isn't an on/off switch. It's a threshold per type of question, set by the cost of an error, the cost of a review and the AI's measured accuracy.
AI strategyWhere should you invest in AI first?
Nearly nine in ten companies surveyed already use AI, and most still see no effect on the bottom line. An effort × impact matrix for choosing what to do first, with a tool to build your own.
Models that decideWhich threshold should your propensity model use?
When a model with good metrics doesn't pay off, the problem is often a number nobody looks at: the threshold. How to set it based on the cost of an error, calibration, and your team's capacity.